KitKat and Common Sense

What AI can and cannot do

You’re sitting in your car. Parked curbside. Engine running. You see a cat “crouch down in front of the right front tire.” It goes out of sight. Then a woman approaches, bends down in front of the car and is obviously looking for the cat. “KitKat come here,” she says.

What do you do? You get out of the car and ask if there’s anything you can do to help.

The robotaxi? It pulls away from the curb and runs over the cat. And that’s all for KitKat the bodega cat, the feline martyr to autonomous progress.

Despite billions of miles run in simulations, and untold number of miles on the road with test drivers and without, the Waymo robotaxi is at a complete loss and utterly unequipped for the one decent, human, sensible act in that situation:

Stop what you’re doing and get out of the car.

The human brain, writes Angus Fletcher in his book Primal Intelligence — You Are Smarter Than You Know, “isn’t brilliant at one mental task; it’s satisfactory at lots, explaining how AI can be so much smarter than the brain yet so much dumber. AI takes one feature of intelligence-- logic -- and accelerates it. As long as life calls for math, AI crushes humans. it’s the king of big data choices. The moment, though, that life requires commonsense or imagination, AI tumbles off its throne.”

The unhappy demise of KitKat in late October reminded me of another story from childhood. I was riding with my family at night on a busy city loop road four lanes wide. An old man stood at an intersection, frozen. He looked confused and frightened. Would he have enough time to cross when the light went his way? Would anyone see him if he got caught in the middle of the road when the lights turned against him?

So my father stopped the car in the road and got out. He walked the old many across, waving for traffic to stop. The cars saw this little drama and yielded. The old man got across and my father jogged back to our car, still idling in a traffic lane, and we went on. A simple act of human decency.

What would a robotaxi have done in that situation?

The demise of a single neighborhood cat has generated an astounding amount of outrage from San Francisco locals. Many are demanding that lawmakers crack down on the robot cabs.

The company’s response?

Waymo, according the Guardian, declined to comment on the proposed legislation that would allow counties to decide whether they will permit the operation of autonomous vehicles in a written statement but said that “trust and the safety of the communities we serve is our highest priority.”.

More from a Nov. 5 Guardian report:

KitKat’s death is the latest swell in waves of anti-AI sentiment and concerns over autonomous vehicles, although others point to data showing a firm safety record. While Waymo says it operates around 1,500 cars across the US, exact numbers for San Francisco’s fleet are unclear. Protestors and activists have taken to disabling Waymos by placing traffic cones on their hoods or even setting them on fire. Cruise, another autonomous taxi company, agreed last year to shell out more than $8m in 2024 to a Bay Area woman who was dragged over 20ft of pavement by an autonomous vehicle. Questions have also arisen over how to ticket and discipline autonomous vehicles when they violate traffic laws.

The San Francisco publication Mission Local asked an expert how a super-smart AI vehicle like a Waymo could do something so stupid.

Professor Missy Cummings, the head of George Mason University’s Autonomy and Robotics Center, said what Waymos don’t have compared to human drivers is “object permanence.” A driver, she added, would have likely seen the cat go under the car and understood that it had not disappeared.

“It’s out of sight, out of mind. Human or animal, anything that falls underneath the car cannot be seen,” said Cummings last month. “It’s forgotten about immediately, because [autonomous vehicles] don’t have a memory.”

No memory. No common sense. No imagination. Out of sight, out of mind.

You could almost forgive Waymo for its bland, emotionally sterile, anonymous boilerplate response to KitKat’s demise because, after all, it’s a cat and cars run over cats all the time. But here’s another missive from the Waymo PR department in response to its robotaxis illegally zipping by school buses. Huge bright yellow school buses with flashing read lights all over and loaded with school kids.

“While we are incredibly proud of our strong safety record showing Waymo experiences twelve times fewer injury crashes involving pedestrians than human drivers, holding the highest safety standards means recognizing when our behavior should be better,” Waymo wrote in a statement, saying it would keep analyzing how its vehicles perform and “making necessary fixes as part of our commitment to continuous improvement.”

So Waymo effected a software patch. Yes, just like when your smartphone has trouble with that restaurant reservation app. But the violations continued.

KXAN reported that Austin school officials said “Waymo’s 20th documented school bus violation occurred on Dec. 1. The company told KXAN and the district it completed software updates across its fleet to prevent its vehicles from passing stopped buses on Nov. 17.”

And the same thing is happening in Atlanta.

Only the most credulous technophile could believe the safety claims robotaxi companies are making. The public backlash against driverless cars will only increase with their numbers.

And then what happens when these robotaxis kill a human? Will Waymo issue another boilerplate press release affirming that safety is Job One?

Kill a human? Really? Don’t take my word for it.

Waymo CEO Says Society Is Ready for One of Its Cars to Kill Someone

And yet the the company is bracing for the first time when a Waymo does kill somebody — a moment its CEO says society will accept, in exchange for access to its relatively safer driverless cars.

“We really worry as a company about those days,” said Waymo co-CEO Tekedra Mawakana on Monday during TechCrunch‘s Disrupt summit, as reported by SFGate. “You know, we don’t say ‘whether.’ We say ‘when.’ And we plan for them.”

Yes, the robotaxi developers claim, we’ll all be incredibly safer with autonomous vehicles. Relatively safer. AI is accelerating development! However, we might kill you. But we won’t kill you as often as you kill yourself. We warned you so don’t come griping to us when this happens. And don’t talk to us about school buses!

There, feel safer?

As for me, “I’m gonna go with the humans.”